The message below was originally posted on the HLA Evolved PSG reflector and has just now been brought to attention again. We’re now trying to reach a larger HLA users community, hoping for some more feedback on the two questions below, in order to decide how to continue with the comment in the March 22 HLA PDG meeting.

I must apologize for not being very familiar with the SR FOM, but I think it does not contain a structure equivalent to the Entity Type in DIS and RPR FOM. But maybe you are able to provide feedback on the second question. Thank you in advance for your support!

(Note that the below message is also posted on the DIS / RPR FOM PSG and RPR FOM PDG reflectors. Please feel free to forward it to any HLA user/community you know.)


From: SAC-PSG-HLA-EVOLVED <[log in to unmask]> On Behalf Of Rene Verhage
Sent: Monday, November 06, 2017 9:23 PM
To: [log in to unmask]
Subject: Request for "complex enumerations" use cases

 

This post is a reach out to the HLA users community, in particular the Object Model designers, to find out to which degree there is a need for enumerations other than based on single (unsigned) integer values. The background is a comment that has been raised (COM_43) in the process of the HLA standard renewal, aka "HLA 4". Upon discussing this comment within the PDG, it has been decided to issue a call for use cases in order to get a better understanding of the usefulness of changing the HLA standard in line of the comment.

The reason the comment has been raised is one particular use case, a well-known set of enumerations/enumerators within the DIS and RPR FOM platform simulation community: the Entity Type. For the purpose of clarifying the concept of "complex enumerations" this use case is presented in more detail below. The HLA community is requested for feedback on the following two questions, including examples if possible:

  1. Do you consider it useful if it would be possible to include Entity Type enumerators in your (RPR FOM based) SOM or FOM, e.g. to express the supported set in your federate or the agreed set in the federation?
  2. Can you think of any other "complex enumeration" datatype that you would like to use in your Object Model; any form of fixed record (or array, or variant record datatype) for which it would be useful to specify a (named) set of allowed value combinations for that datatype?

The RPR FOM datatype EntityTypeStruct, based on the DIS Entity Type record, is a fixed record consisting of 7 fields; 6 fields of 1 octet and 1 field of 2 octets in length, thus in total 8 octets (64 bits).

Entity Kind

Domain

Country

Category

Subcategory

Specific

Extra

Apart from each of these individual fields being an enumeration, also the entire structure is an enumeration. Page 60 to 435 of SISO-REF-010 provide a large set of named entity type values, for shared usage within distributed simulation, enabling simulation interoperability. Each of these enumerators is presented (as they are stored in the source data SISO-REF-010.xml) as a set of decimal values for the record fields, e.g. 1.2.225.1.3.1 for "F-16A" (according to the DIS standard, fields not used shall contain the value 0).

In the current version of the HLA OMT it is not possible to capture the enumerator with the individual field values. It would be possible to treat the entire record as an unsigned integer of 64 bits, enabling to model the allowed set of values, to be referenced by their name. However, the SOM or FOM would then not reveal the build-up of the entity type from its individual field values, hiding e.g. that 72621510375964928 represents a US (225) "F-16A". Also, the current HLA standard has no capability to specify a union between the unsigned 64-bit integer and the EntityTypeStruct. To explicitly allow federate developers to apply such a union, e.g. in order to easily filter the entities based on their kind or domain, the existence of a union would need to be captured in a note.

Now imagine that the EntityTypeStruct could be used for the representation datatype of an enumeration. Then the example entity type enumerator could look like this:

Enumerator name

Field name

Value

F-16A

EntityKind

1

Domain

2

CountryCode

225

Category

1

Subcategory

3

Specific

1

Extra

0

With the first question the HLA PDG would like to gather feedback on whether or not such a capability to include the Entity Type enumeration in the SOM or FOM would be appreciated by the RPR FOM users.

With the second question the HLA PDG invites in general all HLA users to consider whether they would appreciate the capability to use other than integer datatypes for enumerations. Do you know of datatypes similar to the EntityTypeStruct for which it would be beneficial to list the commonly agreed value sets in the Object Model? Could you imagine other datatype structures for the enumeration representation datatype; maybe fixed record datatypes including not only integer datatype fields, or even array or variant record datatypes?

Or would such a capability be regarded as more confusing than helpful in an Object Model? Or is it regarded as undesirable since the concept of complex enumerations does not map easily to many programming languages? Do such counterarguments then apply to the capability in general or only to the use of any datatype for the representation, i.e. would it be acceptable to restrict to fixed records with integer fields only?

Please share with us your thoughts and use cases on this topic. All comments are highly appreciated.

The "HLA 4" PDG



To unsubscribe from the SAC-PDG-SRFOM list, click the following link:
https://discussions.sisostds.org/index.htm?SUBED1=SAC-PDG-SRFOM&A=1